Page 1 of 1

Yr. to yr. win change with and w/o dedicated analysts

Posted: Fri Sep 02, 2011 4:49 am
by Crow
I might be missing a few dedicated analysts but I did a quick check of year to year of win change for teams with and without known dedicated analysts (to my knowledge) over the last 2 seasons I think it was 8 improvements and 8 declines for teams with dedicated analysts and 9 improvements and 5 declines for those without.

11 of the 16 with analyst were in the playoffs though, while only 5 of the 14 without were in.

Can't say for sure how these stats would look for the teams with analysts if they weren't there.

Next season might be even more challenging from an improvement / decline standpoint for the team with analysts. I would not be that surprised if only 5 of 16 improved their win total next season. But I still expect 11 to make the playoffs.

Not sure how many of those teams without analysts will improve next season but I kinda doubt any not in the playoffs last season will improve enough to make the playoffs. More likely it will be a team with an analyst that moves up next season, if any do, in my opinion. Some teams without known analysts might knock harder on the door or get thru in 2012-3 or 2013-4.

Will be interesting to see what happens, eventually.

There is probably more that could be done to analyze upward and downward movement with and without dedicated analysts over a longer period of time: looking at offensive and defensive efficiency improvements and declines separately, whether the improvements at individual level tend to show up more with star or role players, young vs old, big minute lineups or more diffuse, etc.

The changing ratio between actual wins and expected wins based on point differential might be worth looking though it may be heavily influenced by random noise. It could be something of a measure of chemistry / coaching / analytic impact. But not separately for these actors, at least not unless without additional work like a RAPM run with coaches and with & without known analysts. That could be interesting. I'd think that could work if the time period was long enough to provide enough with and without analyst datapoints just as estimates were made for coaches. Not sure what the size of the errors were with coaches though.

One could conceivably also measure with and without heavy use of Synergy data or advanced team collection of "new / better data" by whatever means if you had the scoop on what teams were using and enough without and with datapoints for the data source additions. Theoretically you could get some readings of short-term impacts of dedicated analysts relative to the value added later by the next wave of data enhancements from an even more complex RAPM model and perhaps have something to chew on regarding the value of analysts vs better data debate. Not sure if there would be any or many datapoints with the data additions without dedicated analysts but perhaps there would be a few.

Re: Yr. to yr. win change with and w/o dedicated analysts

Posted: Fri Sep 02, 2011 11:52 am
by Mike G
Crow wrote:... over the last 2 seasons I think it was 8 improvements and 8 declines for teams with dedicated analysts and 9 improvements and 5 declines for those without.

11 of the 16 with analyst were in the playoffs though, while only 5 of the 14 without were in.
...
Maybe an analyst, like many a coach, has an effective shelf life in a given place. A coach is sometimes let go when it's deemed he doesn't have much more to teach, or the players have stopped absorbing it.

There may also be an optimum interval of time for absorption of a given analyst's suggestions. It's got to be unique to each situation. But if you had a spreadsheet of year to year, you might find some trend; for example, if teams rarely get worse after 2-3 years with analysis help.

If 16 analysts work for 11 playoff teams vs 5 sub-playoff teams, and they can't all keep improving, then maintaining success (8+ and 8-) is pretty good.

Re: Yr. to yr. win change with and w/o dedicated analysts

Posted: Sat Sep 03, 2011 12:14 am
by Crow
Core player talent has a life cycle too, related to aging and relative value delivered for the money within a budget constraint.

Teams with dedicated analysts to date have mostly tended to add them while on the rise or from a starting point of existing team strength and the likelihood of more time on a high plateau. Judging the impact of analysts based on just that early set of raw data might tend to be too rosy, too favorable. Now more are finding themselves in the declining part of the team performance life cycle. It will be interesting to try to see what differential impacts can be observed in those phases between with and without analysts.

By showing the contrast between the 2 measures of performance I wanted to note that performance on one indicator is not really enough to fully evaluate the impact of analysts.

I could compile the raw win change data and playoff participation for the last 5+ years and that would probably tell a somewhat richer story but a RAPM style analysis would be more intriguing I’d guess.